University of Birmingham
Comparative analysis of banana wastebioengineering into animal feeds and fertilizersNannyonga, Stella; Mantzouridou, Fani; Naziri, Eleni; Goode, Kylee; Fryer, Peter; Robbins,PhillipDOI:10.1016/j.biteb.2018.04.008
License:Creative Commons: Attribution-NonCommercial-NoDerivs (CC BY-NC-ND)
Document VersionPeer reviewed version
Citation for published version (Harvard):Nannyonga, S, Mantzouridou, F, Naziri, E, Goode, K, Fryer, P & Robbins, P 2018, 'Comparative analysis ofbanana waste bioengineering into animal feeds and fertilizers', Bioresource Technology Reports, vol. 2, pp. 107-114. https://doi.org/10.1016/j.biteb.2018.04.008
Link to publication on Research at Birmingham portal
General rightsUnless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or thecopyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposespermitted by law.
•Users may freely distribute the URL that is used to identify this publication.•Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of privatestudy or non-commercial research.•User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?)•Users may not further distribute the material nor use it for the purposes of commercial gain.
Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document.
When citing, please reference the published version.
Take down policyWhile the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has beenuploaded in error or has been deemed to be commercially or otherwise sensitive.
If you believe that this is the case for this document, please contact [email protected] providing details and we will remove access tothe work immediately and investigate.
Download date: 11. May. 2020
Accepted Manuscript
Comparative analysis of banana waste bioprocessing into animalfeeds and fertilizers
Stella Nannyonga, Fani Mantzouridou, Eleni Naziri, KyleeGoode, Peter Fryer, Phillip Robbins
PII: S2589-014X(18)30030-6DOI: doi:10.1016/j.biteb.2018.04.008Reference: BITEB 29
To appear in:
Received date: 1 April 2018Revised date: 19 April 2018Accepted date: 22 April 2018
Please cite this article as: Stella Nannyonga, Fani Mantzouridou, Eleni Naziri, KyleeGoode, Peter Fryer, Phillip Robbins , Comparative analysis of banana waste bioprocessinginto animal feeds and fertilizers. The address for the corresponding author was capturedas affiliation for all authors. Please check if appropriate. Biteb(2017), doi:10.1016/j.biteb.2018.04.008
This is a PDF file of an unedited manuscript that has been accepted for publication. Asa service to our customers we are providing this early version of the manuscript. Themanuscript will undergo copyediting, typesetting, and review of the resulting proof beforeit is published in its final form. Please note that during the production process errors maybe discovered which could affect the content, and all legal disclaimers that apply to thejournal pertain.
ACCEP
TED M
ANUSC
RIPT
1
Comparative analysis of banana waste bioprocessing into animal feeds and fertilizers
Stella Nannyongaa*, Fani Mantzouridoub, Eleni Nazirib, Kylee Goodea, Peter Fryera and Phillip Robbinsa*
aSchool Of Chemical and Biochemical Engineering University of Birmingham, B15 2TT Edgbaston Birmingham United Kingdom
bFood Chemistry Laboratory University of Aristotle Thessaloniki Greece
Email: [email protected]
Abstract:
The efficiency of a bioprocess in terms of cost of production versus product output determines it pragmatism in the field of Bioengineering. Solid state fermentation (SSF) and anaerobic digestion (AD) were compared for banana waste bioprocessing using saccharomyces cerevisiae. The wastes were pretreated by alkaline-delignification and thermal pre-treatment prior SSF and AD respectively. Optimization of the operational parameters for yeast growth kinetics was done using Chapman, Bergter and Andrew models. From the results generally, pH 3.5 to 5.8 and temperature 22 to 30⁰C were optimum for both bioprocesses. Comparatively, SSF was a more economic and efficient banana waste bioprocess than AD. The protein content increased by approximately 7.9% after SSF and by 6.7% after AD. And the lipid contents increased by 5.9 and 5.4%, while the mineral contents increased by 6.3 and 7.5% both after SSF and AD respectively. This suggested the possible of use of the upgraded wastes as animal feed supplements and nutrient rich fertilizers.
Key words: banana wastes: nutrient rich fertilizers: animal feed supplements
1. Introduction
Over 102 million tons of bananas are produced annually, where 35% of each fruit is a peel,
which produces approximately 36 million tons of waste (Vu et al., 2018). Also in various
countries like Uganda, India, Brazil and Nigeria were bananas are consumed as food,
approximately 70 million tons have also been documented as wasted after ripening in peak
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
2
seasons (Fonsah and Amin 2017; Albertson 2016; Sabiiti et al., 2016). Several other agricultural
wastes like fruit pomace, wheat bran, coffe husks, banana wastes and sugarcane bagasse
among others are also generated in million tons annually (Miller, 2017; Ros et al., 2017). The
release of greenhouse gases to the atmosphere from waste incineration and open field
dumping sites, and the depletion of non-renewable energy resources has prompted the
advancement into environmental friendly bio-processing of agricultural wastes (Yang et al.,
2017; Meng et al., 2017). The increase in the world population which prompts more industrial
and agricultural waste production boosted the bioprocessing of wastes through solid state
fermentation and anaerobic digestion (Akobi et al, 2016; Tao et al., 2017). Solid state
fermentation (SSF) and anaerobic digestion (AD) have been successfully used to upgrade wastes
into industrial enzymes, adsorbents, biofuels, fertilizers, animal feeds among others (Dennehy
et al. 2018 ; Ahmad and Danish 2018). However, also these bio-processing advancements have
remained more of ‘research-findings’ than ‘field applications’ as they fail to justify the feasibility
of the processes in terms of efficiency and costs of production (Colón et al. 2017). These
hiccups have been addressed in several other scientific studies of process optimization and
mathematical modelling (Wang et al., 2014). Optimization of solid state fermentation has been
approached through using mixed culture versus pure populations, optimization of operational
parameters like water activity, pH and temperature, optimization of solid waste selection used
among others (Canabarro et al., 2017; Rodríguez et al., 2017). While for anaerobic digestion the
optimization of hydraulic retention time, effect of inhibitory substances like volatile fatty acids,
batch system versus continuous digestion, packing of the beds among others have been
documented (Ward et al., 2014; Miller, 2017; Ros et al., 2017). However a few studies have
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
3
attempted to compare different bioprocesses while using similar operational conditions and
waste substrates as a pathway to optimize the degree of biodegradability. For example Park et
al., 2011 reported solid state anaerobic digestion of organic wastes for methane production and
Liew et al., 2012 reported solid state anaerobic digestion of lignocellulosic biomass for
production of methane. These studies merged solid state fermentation and anaerobic digestion
processes and termed it ‘solid-state anaerobic digestion’. In this study these processes were
studied independently but under similar operational conditions. Secondly, mathematical
models were employed to describe the microbial growth kinetics of S. cerevisiae and also to
ascertain the optimum biodegradation conditions of waste bananas. S. cerevisiae has been
extensively documented in fermentations for ethanol production (Margeot et al., 2009; Matias
et a., 2015; Canabarro et al., 2017), without focusing on the value of the fermented solid
residue.
Mathematical models have been employed during SSF and AD to describe microbial growth.
For example the Gompertz model has been used in SSF to describe the sigmoid growth of
microorganisms in infinite resources (Gompertz 1825); Baranyi model has been used
predictively to describe non-isothermal and isothermal microbial growth curves (Baranyi et al.,
1996); the GinaFit model has been used successfully to describe the survivor behavior of
microbial populations (Versyck et al., 1999). The square root model has also been employed to
highlight the effect of temperature and pH on the microbial growth rates during fermentation
(Smith-Simpson et al., 2005). For this study Bergter Andrew and Chapman models were used to
describe yeast growth during SSF and AD. Andrew and Bergter models have been used in
kinetic studies for anaerobic digestion of lignocellulosic wastes and other organic complex
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
4
substrates (Yang et al. 2017) but have not been documented for yeast growth during SSF and
AD. Also the Chapman model has been used to describe microbial growth in infinite resource
environments but not during waste bioprocessing.
Pre-treatment of organo-cellulosic compounds is usually done to dissociate the complex
carbon-hydrogen bonds and increase the substrate-microorganism contact ratio (Cai et al.
2016). Thermal pretreatment is a common method employed in SSF which also achieves a
sterile-state of the solid substrate (Ghanem et al., 2000). Alkaline delignification involves use of
alkaline solutions like sodium or potassium hydroxide at varying concentrations to solubilize
lignin. Lignin has been reported to be digested by a few microorganisms like Aspergillus Niger
(Hinojosa et al. 2017). For this study these pre-treatment methods were employed to enhance
the digestibility of banana wastes using saccharomyces cerevisiae. Several studies have been
documented on the successful use of saccharomyces cerevisiae during bioprocessing (Wang
and Witarsa, 2016; Lu et al., 2017). Therefore the aim of the study was to compare alkaline
delignified SSF and thermal pre-treated AD of banana wastes for possible use as animal feeds or
nutrient-rich fertilizer.
2. Materials and Methods
Musa acuminata ‘Grand Nain’, AAA Group bananas were purchased green from Birmingham UK
fruit stores e.g. Aldi and Tesco. The samples were sorted and cleaned to remove the damaged
fingers. They were stored in an incubator at 22⁰C for 9 days up to the waste critical stage
(Nannyonga et al., 2016). Whole fruits (peel and pulp) samples were homogenized by
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
5
mechanical chopping and blending into smaller particles of approximately 1.5-2.0 mm for AD
and SSF.
2.1 Alkaline delignification and solid state fermentation experimental setup
SSF experimental setup was carried out according to (Mantzouridou et al., 2015) with some
modification as detailed below. 30g of homogenized wastes (wet weight) were fermented in
300ml flat bottomed flasks for all the experiments (1:10 w/v). Individual samples were adjusted
to initial pH levels of 4.5, 5.6 and 6.2 using a citrate-phosphate buffer containing 0.2M dibasic
sodium phosphate and 0.1M citric acid and the final pH adjusted using a sensitive meter
(±0.01).
Alkaline delignification pretreatment of banana wastes was done according to (Dahunsi et al.,
2016); Wet state sodium hydroxide at varying concentrations of 4, 8, 12 and 16% (w/w) was
used. The temperature was controlled in a water bath during the pre-treatment. Sodium
hydroxide (NaOH) concentration was calculated as below;
𝑁𝑎𝑂𝐻𝑐𝑜𝑛𝑐 (%) = 𝑚𝑁𝑎𝑂𝐻
𝑑𝑟𝑦 𝑤𝑒𝑖𝑔ℎ𝑡 𝑜𝑓 𝑊𝐵𝐹 𝑥 100 [1]
Where m = number of moles and BW= whole banana fruit.
The samples were inoculated with 1.0 × 105 to 1x107cfu/ml before fermentation and incubated
at different temperatures (22, 28 and 34⁰C) for a period of 7 days. Sampling was done every
24hrs for further analysis. SSF experiments were carried out in triplicates with blank un-
inoculated samples run along assays. Results were expressed in log10 cfu/g of the substrate or
cell concentrations.
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
6
2.2 Thermal pre-treatment and Batch-system anaerobic digestion setup
Batch-system setup according to (Borowski 2015) was employed for banana waste anaerobic
digestion with some modifications as detailed below. The loading volumes were scaled down to
laboratory scale with the ratio of inoculum; banana waste (w/w) of 2.5; 30, 5; 30 and 7.5; 30.
100ml serum bottles were used for digestion of 30g waste bananas under varying initial pH of
4.6, 5.8 and 6.6 respectively. The adjustment of pH was done as detailed in section 2.1. The
samples were thermal pre-treated for 15 minutes at 120⁰C prior inoculation (Mantzouridou et
al. 2015). After cooling the feedstock was inoculated once with 1x105 - 1x107cfu/ml at the
commencement of the digestion process. Nitrogen gas was flushed into the serum bottles in
order to acquire the anaerobic conditions. Biomerieux, France anaerobic atmosphere
indicators and a litmus test were used to ascertain the total absence of oxygen. Anaerobic
digestion was carried out at three mesophilic temperatures which were 25, 30 and 37⁰C and for
a retention time of 10 days. The experiments were carried out in triplicates with un-inoculated
blank samples run along assays. Sampling for further analysis was done in a pattern of 1day, 2
days and 10 days respectively. Serum bottles that were opened for sampling were subjected to
similar oxygen depleting procedures before continuation with further anaerobic digestion. All
kinetic results are expressed as log10 cfu/g of the cell concentration or substrate.
2.3 Inoculum stock preparation
Saccharomyces cerevisiae commercial strain (Saf-instant) was collected from the University of
Birmingham Biochemical Engineering stock bank. The cultures were revived in potato dextrose
broth (PDB) with shaking overnight at 120 rpm at 24°C. For long term storage, the stock
cultures were aliquoted in yeast extract broth (YEB) with a composition of (glucose, 10.0g;
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
7
peptone, 5.0g; yeast extract, 3.0 g; and malt agar, 3.0 g) supplemented with 20% glycerol and
stored in an ultralow-temperature freezer at -80⁰C (Cerda etal., 2017). The inoculum culture
was prepared by adding 2ml of the cell suspension into 100ml of YMB (Yeast Malt Broth) from
Sigma Aldrich consisting (g/l); Dextrose, 10g; Yeast Extract, 3g; Malt agar, 3g and Peptic Digest
of Animal Tissue, 5g. The final pH of the solution was 6.2±0.01 at 25⁰C. The culture was then
incubated at 30⁰C with shaking at 200rpm for 24hrs. Yeast cells were centrifuged in 50ml BD
Falcon, Franklin Lakes, NJ, conical tubes at 8000xg for 10min. They were washed and re-
suspended in PBS saline solution. The PBS composition (g/l) was sodium chloride, 8.0;
potassium chloride, 0.2; disodium hydrogen phosphate, 1.15; and potassium di-hydrogen
phosphate, 0.2. The cell concentration was adjusted to 1x106 cfu/ml with sterile YM broth as
determined turbid-metrically using a spectrophotometer at 600 nm (Cavaleiro et al. 2017).
2.4 Analytical methods
Total suspended solids and volatile solids
The volatile solids were determined by incubating the sample in a furnace at 550⁰C for two
hours (Aggelopoulos et al. 2014). And total solids were determined by heating the sample at
105⁰C until a constant weight (Kalia et al., 2000).
Soluble sugars and COD
The soluble sugars were extracted using hot ethanol and analyzed using UV-spectrophotometry
at 490nm against a blank and glucose standard (Michel Dubois et al., 1956). The COD analysis
were done according to APHA standard method (APHA 2005) A digital Hatch reactor (DRB200,
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
8
15x16 mm vial well and 115vac) and a COD cell test kit were used to measure COD of the
digestate with a detection range of 100 to 1500mg/l (De-Castro and Sato, 2014).
Crude proteins and lipids
Soxhlet extraction was employed for lipids analysis using petroleum ether and Kjeldahl
digestion was used for determination of the protein content (Cordova et al. 1998).
Ash/Minerals, moisture content and total dietary fibre (TDF)
The ash, mineral contents were determined by heating the sample in 525⁰C and the moisture
content was determined by heating at 105⁰C respectively until a constant weight (Deswal et al.,
2011). Total dietary fibers were determined by Happi Emaga et al., 2007 method using
Megazyme kits.
Resistant and non-resistant starch
Total starch (resistant and non-resistant) was determined by (Yaro, 2016) method. Banana
samples were lyophilised and grounded into powder before the analysis using Megazyme
(Megazyme International Ireland A98 YV29) kits (Karthikeyan A, 2010).
Microbial kinetics and modelling
Mathematical models have been employed in SSF and AD to describe the mechanism of growth
for microbial populations under varying operational conditions. Chapman and Richard’s model
is an old model used to describe the growth of microorganisms in infinite resources (Versyck et
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
9
al. 1999) and has not been documented in use during SSF and AD. Other models used in this
study to describe yeast kinetics were Andrew and Bergter models (Wang and Witarsa, 2016).
Chapman-Richards model (1973) is represented as;
𝑝𝑡 = log|𝑛𝑡
𝑛𝑜⁄ | = 𝑋𝑎 |1 − (1 − 𝛽3)𝑒𝑥𝑝
𝛽3
𝛽3𝛽3−1𝜇
𝑋𝑎(𝜆𝑧 − 𝑡) + 𝛽3|
1
1− 𝛽3
[1]
Where n0 = the initial population size and nt = the population size at time, t, μ is the growth rate
[(β0β2𝛽3
𝛽31−𝛽3 )], which is also the gradient of the tangent line, Xa = 𝛽0 are asymptote functions
on the logarithm scale for maximum growth, and 𝜆𝑧is the lag phase which is the x-axis intercept
of the tangent line.
The lag phase is mathematically represented as;
𝜆 = 𝛽0(1− 𝛽1)
11− 𝛽3− 𝛽0𝛽3
11− 𝛽3+ 𝜇
𝑙𝑜𝑔(𝛽1
1− 𝛽3)
𝛽2
𝜇 [2]
The Andrew’s model is represented as;
µ = µ𝑚𝑎𝑥 ∗ 𝑆
𝐾𝑠+𝑆+ 𝑆2
𝐾𝑖
[3]
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
10
And Bergter model as;
µ = µ𝑚𝑎𝑥 ∗ 𝑆
𝐾𝑠+𝑆 ∗ |1 − 𝑒𝑥𝑝 [
−𝑡
𝜆]| [4]
Where µ is the growth rate (h-1), µmax is the maximum growth rate (h-1), 𝜆 is the lag phase (h-1), t is the
acceleration time, Ks is the substrate concentration constant (mg/l-1), S is the substrate concentration
and Ki the Inhibition coefficient (mg/l-1).
3. Results and Discussions
Delignified SSF and thermal pre-treated AD of banana wastes using saccharomyces cerevisiae
was observed under varying inoculum sizes, pH levels and temperatures to ascertain optimum
operational conditions. Under these optimum conditions chemical characterization was done
and compared to ascertain the degree of digestability of the waste and the possible use of the
upgraded waste. Triplicate runs were carried out under optimum conditions. Mathematical
models were applied to describe and compare the growth kinetics of yeast.
3.1 Comparison of the initial inoculum concentration during alkaline delignified fermentation
and thermal pre-treated anaerobic digestion
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
11
Inoculum size, pH and temperature were used to study yeast growth kinetics during banana
waste bioprocessing. Mathematical models were employed to describe the biological
parameters of the kinetics i.e. lag phases, substrate constants and growth rates.
The optimum initial inoculum concentrations required to digest the same amount of banana
wastes under varying pH and temperature were determined during SSF and AD. Initial inoculum
sizes of 1.0 x 105 and 1.0 x 106 were used for this study with 30g of banana wastes Figure 1. The
selection of the inoculum sizes depended on literature and single test experiments
(Mantzouridou et al. 2015).
From the SSF results, the initial inoculum size of 1x105cfu/g increased to a maximum microbial
population of log108cfu/g at approximately 70 hours. While the initial inoculum 1x106cfu/g
increased to a higher maximum microbial population of log109cfu/g also at 70 hours, however it
can be seen (Figure 1) that the peak population is followed by a tremendous decline (death)
phase. This therefore rendered the initial inoculum size of 105 more optimum as compared to
106 for delignified SSF of banana wastes.
Considering AD results, the initial size of 1x106cfu/g increased to a higher microbial populations’
of log107cfu/g as compared to initial sizes of 1x105cfu/g. The former sustained the maximum
microbial population for approximately 70 hours while the latter sustained the maximum
microbial population for approximately 50 hours. This also indicates that the initial size of 106
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
12
cfu/g is more optimum as the maximum population was sustained longer as compared to the
initial inoculum size of 105cfu/g for thermal pre-treated anaerobic digestion of banana wastes.
The results highlighted the significance of inoculum optimization, it should be noted that the
size of inoculum to effect a bioprocess is necessary to optimize as very low concentrations
results in long digestion cycles which require high energy and costs of production. While high
initial inoculums also result in high populations during the initialization stage of the digestion
process leading to high death rates before commencement of digestion (Batstone et al., 2015).
For comparative evaluation, delignified SSF require a lower inoculum size as compared to
thermal pre-treated anaerobic digestion for the same amount of banana waste.
3.2 Effect of pH on the growth kinetics of saccharomyces cerevisiae during delignified SSF and
thermal pre-treated AD of banana wastes
The effect of pH on S. cerevisiae growth kinetics was studied to ascertain the optimum
conditions during delignified SSF and thermal pre-treated AD. Initial pH levels of 4.5, 4.6, 5.6,
5.8, 6.2 and 6.6 were considered for this study (Figure 2 and 3).
The growth kinetics of yeast during delignified SSF under varying pH levels indicated pH 4.5 as
optimum, followed by pH 5.6 and lastly pH 6.2 (Figure 2). The maximum microbial population
obtained was log108cfu/g and was sustained for approximately 90 hours under pH 4.5
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
13
fermentations. Low microbial populations with a maximum of log106 and log104cfu/g were
shown by pH 5.6 and 6.2 respectively.
The thermal pre-treated AD (Figure 3) also indicated similar variations as with SSF, with pH 4.6
as optimum for yeast growth as compared to pH 5.8 and 6.6. The maximum microbial
population at pH 4.6 was log107cfu/g, while the maximum for pH 5.8 and 6.6 were log106 and
log105cfu/g respectively.
Comparative evaluation of yeast growth kinetics under varying pH during SSF and AD indicated
higher growth during the former than the latter. Also less time was required to reach the
maximum microbial population under all pH levels. For example considering the optimum pH
(4.5-4.6), it took about 70 hours to reach a maximum microbial population during SSF and
about 120 hours for AD.
Longer lag phase (Figure 3) were also shown during AD as compared to SSF (Figure 2) under
similar conditions.
3.3 Effect of temperature on the growth kinetics of saccharomyces cerevisiae during
delignified SSF and thermal-pretreated AD
The effect of temperature on the growth kinetics of yeast during SSF and AD was determined.
Temperatures 22, 25, 28, 30, 34 and 3⁰7C were considered for this study (Figure 4 and 5).
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
14
During delignified SSF, temperature 22 and 28⁰C indicated similar patterns of growth with
higher maximum microbial populations in the former with log108cfu/g than the latter with
log107cfu/g. Low microbial populations were shown by temperature 34⁰C with a maximum of
log104cfu/g.
Microbial growth studies during AD under varying temperatures indicated that operation under
25 and 30⁰C could sustain similar maximum microbial populations of up to log107cfu/g
however, under 30⁰C the population declines just after 20 hours.
Optimization of pH and temperatures during SSF and AD did not only reduce the lag phases but
also increased the growth rates (Figure 2-5). It was concluded that the optimum pH for yeast
during banana waste bioprocessing was between 4.5-4.6, and temperature 22-30⁰C.
Optimisation of operational parameters like pH and temperatures have been documented to
increase the production of biofuels and industrial enzymes during agro-waste bioprocessing
(Lee et al., 2017; Leite et al., 2016; Wang et al., 2018).
From the results it was noted that similar optimum ranges of temperatures i.e. between 22 to
30⁰C were show for banana waste AD and SSF using S. cerevisiea. Comparatively, higher
microbial popualations and less time to reach the maximum population were required during
delignified SSF than thermal pre-treated AD.
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
15
3.4 Mathematical modelling of the growth kinetics of S. cerevisiae during delignified SSF and
thermal pre-treated AD
Chapman, Bergter and Andrew’s models were employed to describe yeast growth kinetics
during SSF and AD of banana wastes. Chapman model results were derived from the
mathematical functions of the sigmoid growth curve and they were compared to the derived
kinetic parameters of Bergter and Andrews model (Table 1). Temperature 22 and 25⁰C were
used as reference for delignified SSF and thermal pre-treated AD respectively under varying pH
levels. The kinetic derived parameters from Bergter model were lag phase (λ), maximum
growth rate (µmax) and substrate constant (Ks) while from Andrew’s model, the derived
parameters were maximum growth rate (µmax) and substrate constant (Ks) and inhibitor
constant (Ki).
Chapman model results for SSF indicated pH 4.5 as optimum with the lowest lag phase of
8.353±0.030 and highest growth rate of 2.391±0.015 as compared to other pH ranges. Also for
AD pH 4.5 indicated the optimum conditions with the lowest lag phase of 10.024±0.011, and
high growth rate of 1.024±0.011. From the selected pH levels for both processes it was noted
that pH 4.5 to 4.6 were optimum for yeast growth during banana waste bio-processing. Similar
trends were followed for maximum growth rates (μmax) of yeast under varying pH levels. The
maximum rate was observed during delignified SSF as 2.580±0.021 as compared to thermal pre-
treated AD with 1.106±0.031. Lower pH ranges i.e. 3.5 to 5.6 were considered optimum for SSF
and AD of waste bananas using Saccharomyces cerevisiae.
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
16
The comparative results for the mathematical models indicate close values between Chapman
and Bergter models as compared to Andrew’s model (Table 2). For example lag phases for
Chapman model during pH digestions were 10.024±0.015 and 10.024±0.010 for Bergter model.
The RMSE for predicted model results and actual experiments were very low with 0.011 for
Bergter model.
Also the kinetic derived parameters for delignified SSF and thermal pretreated AD indicate
lower lag phases and maximum growth rates. For example the predictive maximum growth rate
during SSF was 2.581±0.015 and that of AD was 1.098±0.020 using Bergter and Andrew models
respectively. Also Chapman model results indicated delignified SSF as a more efficient process
as compared to thermal pre-treated AD, in terms of shorter lag phases and faster growth rates
under similar operating conditions.
Mathematical model predictive results highlighted the ability of Chapman and Bergter models
to describe yeast growth as compared to Andrew’s model (Table 2). The results also indicated
higher efficiency in terms of less time, inoculum size, and maximum population registered
during SSF as compared to AD. The greatest hiccup of agro waste bioprocessing is higher energy
and costs of production as compared to the value of the product generated (Jena et al. 2017;
Puyol et al., 2017). Long cycles can be minimized by reducing the lag phases by using
mathematical models to determine optimum operational conditions before large scale bio-
processing (Rasmussen et al., 2010; Ruan et al., 2014).
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
17
3.5 Comparative chemical characterization of banana wastes during delignified SSF and
thermal pre-treated AD
Chemical analysis was done on the upgraded banana waste after delignified SSF and thermal
pre-treated AD, Table 3. The purpose of chemical characterization was to compare the degree
of digestability of waste bananas during SSF and AD in terms of upgrading the wastes into
animal feed supplements and fertilizers. Several parameters were considered and compared
with literature, Table 3. The key parameters to be considered for the above purposes are
proteins, lipids and mineral compositions.
Comparatively, higher protein levels were registered after delignified SSF as compared to
thermal pre-treated AD, the former consisted 10%, while the latter consisted 8.8% from an
initial of before bio-processing 2.1%. Also the crude lipid content followed similar trends with
higher percentages reported during SSF (7.8%) as compared to AD (7.0%) from an initial of
1.6%.
The other chemical compositions for banana wastes after SSF and AD indicate a uniform trend
of higher digestability during the former than the latter. For example the recovered soluble
sugars after SSF were 12.6% and 16% after AD. Higher mineral compositions were registered
after AD as compared to SSF with percentages of 9.8 and 8.6 respectively.
Chemical characterization of the feedstock, highlighted higher proteins and lipids in the
delignified SSF wastes as compared to thermal pre-treated AD wastes. This concluded the
former as nutrient rich animal supplement as compared to the latter. However higher mineral
composition in anaerobic digested wastes as compared to the fermented wastes indicated the
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
18
former as a nutrient rich fertilizer than the latter. This highlighted how comparative evaluation
of bioprocess can aid large scale applications with a more specified purpose.
3.6 Conclusions
Comparative evaluation of delignified SSF and thermal pretreated AD clearly highlighted the
necessity to compare different bioprocesses in order to determine the more efficient for large
scale and commercial applications. Delignified SSF was considered more economical and
feasible as compared to thermal pretreated AD in terms of initial inoculum size, retention time
for bioprocessing, larger spectrum of pH and temperature tolerance for the microbial
population to survive among others. Also higher microbial populations were registered during
SSF which resulted in a more upgraded waste with higher protein, lipids and mineral
compositions. Optimization of operational parameters (inoculum sizes, pH and temperature)
also indicated higher bioprocessing rates under the optimum conditions, with pH 3.5 to 5.8 and
temperature 22 to 30⁰C being optimum for yeast growth. Chemical characterization of the
upgraded wastes for both bioprocesses indicated increase in proteins by 7.9% after SSF and by
6.7% after AD. The lipids increased by 5.9 and 5.4%, while the mineral contents increased by 6.3
and 7.5% both after SSF and AD respectively. This suggested possible use of the upgraded
banana wastes as animal feed supplements or fertilizer. These bioprocessing methods can be
employed to address banana wastage in peak seasons from third world countries like Uganda
and Brazil.
‘E-Supplementary data for this research study can be found in e-version of this paper online’
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
19
Acknowledgement
The authors acknowledge the School of Chemical Engineering University of Birmingham for
funding the study. The researcher is also grateful to God for enabling her pursue her PhD
studies.
References
Aggelopoulos, T. et al., 2014. Solid state fermentation of food waste mixtures for single cell
protein, aroma volatiles and fat production. Food Chemistry, 145, pp.710–716.
Ahmad, T. and Danish, M., 2018. Prospects of banana waste utilization in wastewater
treatment: A review. Journal of Environmental Management, 206, pp.330–348.
Akobi, C. et al., 2016. Single-stage and two-stage anaerobic digestion of extruded lignocellulosic
biomass. Applied Energy, 184, pp.548–559.
Albertson, E., 2016. “Adopting new banana varieties in Uganda: the role of gender and head of
household status.”
APHA, 2005. APHA Standards methods for the examination of water and wastewater. APHA-
AWWA-WEFWashington DC.
Baranyi, J. et al., 1996. Effects of parameterization on the performance of empirical models
used in `predictive microbiology’. Food Microbiology, 13(1), pp.83–91.
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
20
Batstone, D.J. et al., 2015. Mathematical modelling of anaerobic digestion processes:
applications and future needs. Reviews in Environmental Science and Bio/Technology,
14(4), pp.595–613.
Borowski, S., 2015. Temperature-phased anaerobic digestion of the hydromechanically
separated organic fraction of municipal solid waste with sewage sludge. International
Biodeterioration & Biodegradation, 105, pp.106–113.
Cai, D. et al., 2016. Effect of dilute alkaline pretreatment on the conversion of different parts of
corn stalk to fermentable sugars and its application in acetone–butanol–ethanol
fermentation. Bioresource Technology, 211, pp.117–124.
Canabarro, N.I. et al., 2017. Ethanol production by solid-state saccharification and fermentation
in a packed-bed bioreactor. Renewable Energy, 102, pp.9–14.
de Castro, R.J.S. and Sato, H.H., 2014. Production and biochemical characterization of protease
from Aspergillus oryzae: An evaluation of the physical–chemical parameters using
agroindustrial wastes as supports. Biocatalysis and Agricultural Biotechnology, 3(3), pp.20–
25.
Cavaleiro, A.J. et al., 2017. Anaerobic Digestion of Lipid-Rich Waste. In T. J. McGenity, K. N.
Timmis, and B. Nogales, eds. Hydrocarbon and Lipid Microbiology Protocols: Pollution
Mitigation and Waste Treatment Applications. Berlin, Heidelberg:
Cerda, A. et al., 2017. Towards a competitive solid state fermentation: Cellulases production
from coffee husk by sequential batch operation and role of microbial diversity. Science of
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
21
The Total Environment, 589, pp.56–65.
Colón, J. et al., 2017. Analysis of MSW full-scale facilities based on anaerobic digestion and/or
composting using respiration indices as performance indicators. Bioresource Technology,
236, pp.87–96.
Cordova, J. et al., 1998. Lipase production by solid state fermentation of olive cake and sugar
cane bagasse. Journal of Molecular Catalysis B: Enzymatic, 5(1), pp.75–78.
Dahunsi, S.O. et al., 2016. Mesophilic anaerobic co-digestion of poultry dropping and Carica
papaya peels: Modelling and process parameter optimization study. Bioresource
Technology, 216, pp.587–600.
Dennehy, C. et al., 2018. Anaerobic co-digestion of pig manure and food waste; effects on
digestate biosafety, dewaterability, and microbial community dynamics. Waste
Management, 71, pp.532–541.
Deswal, D., Khasa, Y.P. & Kuhad, R.C., 2011. Optimization of cellulase production by a brown rot
fungus Fomitopsis sp. RCK2010 under solid state fermentation. Bioresource Technology,
102(10), pp.6065–6072.
Fonsah, E.G. & Amin, B., 2017. Evaluating Overall Performances of the Banana Industry in West
Bengal State, India. Journal of Food Distribution Research, 48(1).
Ghanem, N.B., Yusef, H.H. & Mahrouse, H.K., 2000. Production of Aspergillus terreus xylanase in
solid-state cultures: application of the Plackett–Burman experimental design to evaluate
nutritional requirements. Bioresource Technology, 73(2), pp.113–121.
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
22
Gompertz, B., 1825. On the Nature of the Function Expressive of the Law of Human Mortality,
and on a New Mode of Determining the Value of Life Contingencies. Philosophical
Transactions of the Royal Society of London, 115, pp.513–583.
Happi Emaga, T. et al., 2007. Effects of the stage of maturation and varieties on the chemical
composition of banana and plantain peels. Food Chemistry, 103(2), pp.590–600.
Hinojosa, U.D. et al., 2017. Solid-State Culture for Lignocellulases Production. In Fermentation
Processes. InTech.
Jena, S.P. et al., 2017. An experimental approach to produce biogas from semi dried banana
leaves. Sustainable Energy Technologies and Assessments, 19, pp.173–178.
Karthikeyan A, S.N., 2010. Citric acid production by Koji fermentation using banana peel as a
novel substrate. Bioresource Technology, 101(14), pp.5552–5556.
Lee, J. et al., 2017. Long-term performance of anaerobic digestion for crop residues containing
heavy metals and response of microbial communities. Waste Management, 59, pp.498–
507.
Leite, P. et al., 2016. Ultrasounds pretreatment of olive pomace to improve xylanase and
cellulase production by solid-state fermentation. Bioresource Technology, 214, pp.737–
746.
Li, Y., Park, S.Y. & Zhu, J., 2011. Solid-state anaerobic digestion for methane production from
organic waste. Renewable and Sustainable Energy Reviews, 15(1), pp.821–826.
Liew, L.N., Shi, J. & Li, Y., 2012. Methane production from solid-state anaerobic digestion
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
23
of lignocellulosic biomass. Biomass and Bioenergy, 46, pp.125–132.
Lu, Y. et al., 2017. Combined effects of fermentation temperature and pH on kinetic changes of
chemical constituents of durian wine fermented with Saccharomyces cerevisiae. Applied
Microbiology and Biotechnology, 101(7), pp.3005–3014.
Mantzouridou, F.T., Paraskevopoulou, A. & Lalou, S., 2015. Yeast flavour production by solid
state fermentation of orange peel waste. Biochemical Engineering Journal, 101, pp.1–8.
Margeot, A. et al., 2009. New improvements for lignocellulosic ethanol. Curr Opin Biotechnol,
20.
Matias, J. et al., 2015. Optimisation of ethanol fermentation of Jerusalem artichoke tuber juice
using simple technology for a decentralised and sustainable ethanol production.
Meng, Y. et al., 2017. Enhancing anaerobic digestion performance of crude lipid in food waste
by enzymatic pretreatment. Bioresource Technology, 224, pp.48–55.
Michel Dubois, K. A. Gilles, J. K. Hamilton, P. A. Rebers, and F.S., 1956. Sugars, Colorimetric
Method for Determination of Substances, and Related. niversity of Minnesota, St. Paul,
Minn.Division of Biochemistry.
Miller, B., 2017. Optimization of anaerobic digestion through the introduction of ruminant
methanogens. In 2017 AAAS Annual Meeting (February 16-20, 2017). aaas.
Nannyonga, S. et al., 2016. Mathematical modelling of color, texture kinetics and sensory
attributes characterisation of ripening bananas for waste critical point determination.
Journal of Food Engineering, 190, pp.205–210.
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
24
Puyol, D. et al., 2017. ZVI Addition in Continuous Anaerobic Digestion Systems Dramatically
Decreases P Recovery Potential: Dynamic Modelling. In G. Mannina, ed. Frontiers in
Wastewater Treatment and Modelling: FICWTM 2017. Cham: Springer International
Publishing, pp. 211–217.
Rasmussen, M.L. et al., 2010. Sequential saccharification of corn fiber and ethanol production
by the brown rot fungus Gloeophyllum trabeum. Bioresource Technology, 101(10),
pp.3526–3533.
Rodríguez de Olmos, A., Correa Deza, M.A. & Garro, M.S., 2017. Selected lactobacilli and
bifidobacteria development in solid state fermentation using soybean paste. Revista
Argentina de Microbiología, 49(1), pp.62–69.
Ros, M. et al., 2017. Mesophilic anaerobic digestion of pig slurry and fruit and vegetable waste:
Dissection of the microbial community structure. Journal of Cleaner Production, 156,
pp.757–765.
Ruan, Z. et al., 2014. Oleaginous fungal lipid fermentation on combined acid- and alkali-
pretreated corn stover hydrolysate for advanced biofuel production. Bioresource
Technology, 163, pp.12–17.
Sabiiti, G., J.M. Ininda, L. Ogallo, F. Opijah, A. Nimusiima, G. Otieno, S.D. Ddumba, J. and
Nanteza and C. Basalirwa., 2016. Empirical relationship between banana yields and climate
variability over Uganda. Journal of Environmental & Agricultural Sciences., 7: 03-13.
Smith-Simpson, S. and Schaffner, D.W., 2005. Development of a Model To Predict Growth of
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
25
Clostridium perfringens in Cooked Beef during Cooling. Journal of Food Protection, 68(2),
pp.336–341.
Tao, B. et al., 2017. Enhancement of microbial density and methane production in advanced
anaerobic digestion of secondary sewage sludge by continuous removal of ammonia.
Bioresource Technology, 232, pp.380–388.
V.C Kalia; VSonakya; NRaizada, A., 2000. Anaerobic digestion of banana stem waste.
Bioresource Technology, 73(2), pp.191–193.
Versyck, K.J. et al., 1999. Introducing optimal experimental design in predictive modeling: A
motivating example. International Journal of Food Microbiology, 51(1), pp.39–51.
Vu, H.T., Scarlett, C.J. & Vuong, Q. V., 2018. Phenolic compounds within banana peel and their
potential uses: A review. Journal of Functional Foods, 40, pp.238–248.
Wang, K. et al., 2014. Anaerobic digestion of food waste for volatile fatty acids (VFAs)
production with different types of inoculum: Effect of pH. Bioresource Technology, 161,
pp.395–401.
Wang, P. et al., 2018. Do microbial communities in an anaerobic bioreactor change with
continuous feeding sludge into a full-scale anaerobic digestion system? Bioresource
Technology, 249, pp.89–98.
Wang, Y. & Witarsa, F., 2016. Application of Contois, Tessier, and first-order kinetics for
modeling and simulation of a composting decomposition process. Bioresource Technology,
220, pp.384–393.
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
26
Ward, A.J., Lewis, D.M. & Green, F.B., 2014. Anaerobic digestion of algae biomass: A review.
Algal Research, 5, pp.204–214.
Yang, J. et al., 2017. Estimation of kinetic parameters of an anaerobic digestion model using
particle swarm optimization. Biochemical Engineering Journal, 120, pp.25–32.
Yaro, M.N., 2016. Proximate composition of banana leaves and estimation of fungal
degradationproducts of the leaves at mesophilic temperature. World WIde Journal of
MUltidisciplinary Resaearch and Developmnet, 2(5), pp.31–35.
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
27
Figure 1: Growth kinetics of S. cerevisiae during de-lignified SSF and thermal pre-treated AD at varying initial inoculum sizes
Figure 2: Growth kinetics of S. cerevisiae during de-lignified SSF under varying pH levels
Figure 3: Variation of S. cerevisiae growth kinetics during thermal pre-treated AD under varying pH levels
Figure 4: Variation of S. cerevisiae growth kinetics during de-lignified SSF under varying temperatures
Figure 5: Variation of S. cerevisiae growth kinetics during thermal pre-treated AD under varying temperatures
Table 1: Chapman Richards model results of S. cerevisiae growth kinetics during de-lignified SSF and thermal pre-treated AD
Delignified SSF (ref, 22⁰C)
Thermal pre-treated AD ( ref, 25⁰C)
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
28
pH λ (h) µ (h) μmax (h) pH λ (h) µ(h) μmax (h)
3.5
8.92±0.125b 1.8`±0.073b
c 2.16±0.082a
b 4.6
10.02±0.011a
b 1.02±0.012a 1.11±0.031a
b 4.5
8.35±0.032a 2.39±0.015c 2.58±0.021b 5.8
12.71±0.104c 0.94±0.026b
d 1.02±0.012c
d 5.6
9.85±0.082d 1.17±0.030a 1.39±0.017c
d 6.6
13.91±0.217b 0.62±0.031a
b 0.98±0.043b
c 6.2
10.91±0.264c
1.01±0.105c
d 1.25±0.019a 7.
0 14.11±0.328a
c 0.59±0.006c 0.88±0.084b
d Where λ is the lag phase, μ = growth rate, µmax = maximum growth rates, values are presented as means ± SEM of 17 individual experiments. Also means with different letters are p> 0.05 significantly different
Table 2: Comparative evaluation of S. cerevisiae growth kinetics using the derived parameters of Bergter and Andrew versus Chapman model during SSF and AD
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
29
Model λ (h)
µmax (h)
Ks (mg/l-1)
Ki (mg/l-1)
RMSE
SSE
R2
Delignified SSF (ref, 4.5)
Chapman-Richards 8.353±0.031ab 2.580±0.022b 0.001 0.010 0.99 Bergter 8.357±0.021bc 2.581±0.015cd 173 0.011 0.112 0.99 Andrew 2.551±0.032ab 125 268 0.025 0.143 0.97 Thermal pre-treated AD (ref, 4.6)
Chapman-Richards 10.024±0.015ab 1.106±0.012ac 0.002 0.101 0.99 Bergter 10.024±0.010bc 1.108±0.020ab 159.6 0.020 0.109 0.98 Andrew 1.098±0.020c 52.6 396 0.031 0.138 0.97
Where λ = lag phase; µmax = maximum growth rate; RMSE = root mean square error; Ks = substrate constant; Ki = inhibitor constants; means (±) SEM for 17 individual runs; means with different letters are p> 0.05 significantly different
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
30
Table 3: Chemical characterization of de-lignified SSF and thermal pre-treated AD banana wastes
Parameter
(g/100g)
Non-bioprocessed banana wastes (%)
Delignified SSF banana wastes (%)
Thermal-pretreated AD banana wastes (%)
pH 4.60±0.017c 4.48±0.011a
Minerals (Ash) 2.37±0.015a 8.685±0.069ab 9.88±0.037b
Total Volatile solids (DW %) 90.02±0.031b 86.26±0.012b
Moisture (FW %) 90.02±0.085b 93.07±0.014ac
Soluble Sugars (DW %) 72.57±0.044ab 12.60±0.211ad 16.03±0.151ab
Protein (DW %) 2.13±0.086bc 10.09±0.121a 8.85±0.035cd
Total solids (FW %) 92.04±0.106cd 88.25±0.021b
Total dietary fiber (DW %) 11.05±0.221c 43.06±0.405b
COD (g O2/litre) 64.72±0.016a 38.73±0.067b
Crude Lipids (DM %) 1.61±0.072b 7.54±0.079c 7.04±0.111bc
Where FW = fresh weight; DW = dry weight; COD = chemical oxygen demand; means (±) SEM for 3 individual experiments; means with different letters are p> 0.05 significantly different
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
31
ACCEPTED MANUSCRIPT
ACCEP
TED M
ANUSC
RIPT
32
Highlights
- Banana waste bioprocessing and mathematical modelling
- Biomass pre-treatment prior solid state fermentation and anaerobic digestion
- Chemical characterisation of upgraded wastes as animal feeds and fertilizers
ACCEPTED MANUSCRIPT
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5